def actions_distance(self, distance):
def entries(i, j):
a1 = self.actions[i]
a2 = self.actions[j]
return distance(a1, a2)
K = len(self.actions)
return construct_matrix((K, K), entries)
def actions_distance(self, distance):
def entries(i, j):
a1 = self.actions[i]
a2 = self.actions[j]
return distance(a1, a2)
K = len(self.actions)
return construct_matrix((K, K), entries)
def get_domain(H, W, noise=0):
""" returns points, distance """
points = []
for i, j in itertools.product(range(H), range(W)):
p = np.array((i, j))
p = p + np.random.randn(2) * noise
points.append(p)
distance = lambda p1, p2: np.linalg.norm(np.array(p1) - np.array(p2))
N = H * W
D = construct_matrix((N, N), lambda i, j: distance(points[i], points[j]))
return points, D
def get_distance_matrix(self, plans, action_distance):
n = len(plans)
D_ij = lambda i, j: self.plan_distance(plans[i], plans[j],
action_distance)
D = construct_matrix((n, n), D_ij)
return plans, D
def get_distance_matrix(self, plans, action_distance):
n = len(plans)
D_ij = lambda i, j: self.plan_distance(plans[i], plans[j], action_distance)
D = construct_matrix((n, n), D_ij)
return plans, D